BDNF and TrkB protein expression is altered in the fetal hippocampus but not cerebellum after chronic prenatal compromise

Experimental Neurology 192 (2005) 265 – 273 www.elsevier.com/locate/yexnr

BDNF and TrkB protein expression is altered in the fetal hippocampus but not cerebellum after chronic prenatal compromise Sandra Dieni, * Sandra Rees Department of Anatomy and Cell Biology, The University of Melbourne, Parkville 3010, Victoria, Australia Received 23 January 2004; revised 17 May 2004; accepted 1 June 2004 Available online 20 July 2004

Abstract This study examines the effects of a chronic prenatal insult on both the expression of brain-derived neurotrophic factor (BDNF) and TrkB proteins and the structural development of the fetal hippocampus and cerebellum. Chronic placental insufficiency (CPI) was induced via unilateral ligation of the uterine artery from midgestation to near term in the pregnant guinea pig. Fetuses were delivered at 60 days of gestation (dg, term approximately 67 dg) and classified as control or growth-restricted (GR) according to established criteria. In hippocampi and cerebella from control (n = 7) and GR (n = 8) fetuses, immunohistochemistry was performed to detect the expression of BDNF and TrkB proteins, and the growth of neuropil and cellular layers was measured in each structure. The growth of neuropil layers was reduced in the dentate gyrus of GR fetuses compared to controls: hippocampi from severely GR fetuses showed a marked reduction in BDNF-IR and an increase in TrkB-IR. The most pronounced effects on neuropil growth were seen in the same fetuses that demonstrated marked alterations in BDNF-IR and TrkB-IR. In the cerebellum, there were significant reductions in the growth of the cellular and neuropil layers; however, BDNF-IR and TrkB-IR were not affected. These results demonstrate that CPI has a widespread effect in retarding process growth in the developing brain, but a differential effect on neurotrophin expression. Changes in BDNF and TrkB expression appear to be associated with the pronounced structural changes in the hippocampi of severely GR fetuses, however, structural abnormalities in the cerebellum were not associated with changes in these proteins; presumably, other factors are involved. D 2004 Elsevier Inc. All rights reserved. Keywords: Development; Neurotrophin; Neuropil; Intrauterine growth restriction; Placental insufficiency; Immunohistochemistry

Introduction Intrauterine growth restriction (IUGR) can result from adverse conditions in utero and has been closely associated with increased morbidity and mortality in the perinatal period (McIntire et al., 1999). Follow-up studies of IUGR infants have demonstrated increased neurodevelopmental (Larroque et al., 2001), learning, and behavioral (Low et al., 1993) problems compared to normally grown infants. A proportion of neonates display evidence of chronic malnutrition (Jones and Parer, 1983), hypoxemia (Jensen et al., 1996), and altered endocrine balance (Jones et al., 1990); such symptoms are characteristic of chronic placental insufficiency (CPI). In a guinea pig model of CPI induced via uterine artery ligation throughout the second half of gestation (Lafeber et al., 1984), * Corresponding author. Department of Medicine, University of Melbourne, Austin Health, Level 7, Lance Townsend Building, Studley Road, Heidelberg 3084, Victoria, Australia. Fax: +61-3-9457 4585. E-mail address: [email protected] (S. Dieni). 0014-4886/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2004.06.003

we have shown that there are reductions in the volume of the hippocampus at term (Mallard et al., 1999), altered dendritic morphology of hippocampal neurons at term (Dieni and Rees, 2003), and fewer CA1 and Purkinje cells in the postnatal hippocampus and cerebellum, respectively (Mallard et al., 2000). Furthermore, in an ovine model of CPI induced by umbilicoplacental embolization in late gestation, we have demonstrated significant abnormalities in the development of the cerebral hemispheres and cerebellum at term (Mallard et al., 1998). Thus, although it is clear that CPI can affect brain development, the events that underlie these changes are still being elucidated. The appropriate growth and connectivity of the fetal brain is highly dependent on the availability of growth factors, including neurotrophins (for review, see McAllister, 2001). In particular, brain-derived neurotrophic factor (BDNF) has been shown to have potent effects on developing neurons, via activation of its high affinity tyrosine kinase receptor, TrkB. BDNF promotes the survival and growth of dentate granule cells (Lowenstein and Arsenault,

266

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

1996) and pyramidal neurons (Bartrup et al., 1997; Ip et al., 1993) in the hippocampus. In addition, BDNF is known to play a role in granule cell migration (Borghesani et al., 2002), foliation, and dendritic outgrowth (Schwartz et al., 1997) during cerebellar development. Recently, it has been reported that the function of cortically expressed BDNF is to support the maintenance of cortical neuronal size and dendritic structure rather than the initial development of these features (Gorski et al., 2003). BDNF and TrkB expression are known to be affected by a broad range of acute brain insults. For example, increases in the synthesis of BDNF or TrkB proteins occur after transient forebrain ischemia in the adult rat brain (Ferrer et al., 1998; Kokaia et al., 1998; Merlio et al., 1993; Takeda et al., 1993) and kainic acid exposure (Rudge et al., 1998), while moderate hypoxic-ischemia in the young rat leads to a dramatic reduction in both BDNF-IR and TrkB-IR in CA1 pyramidal neurons, before cell death (Walton et al., 1999). Thus, while the effects of acute brain insults on the expression of BDNF and TrkB have been examined in the postnatal and mature brain, the influence of a sustained prenatal insult on the expression of these proteins during gestation has not previously been reported. Therefore, using a model that mimics adverse events that can occur in utero, we have generated growth-restricted (GR) guinea pigs to determine whether chronic placental insufficiency affects the expression of BDNF and TrkB proteins in the hippocampus and cerebellum and whether these changes correlate with morphological alterations. The guinea pig was chosen as the model since, as in humans, neurogenesis and a significant proportion of dendritic and axonal proliferation occur in utero (Dieni and Rees, 2002; Mallard et al., 2000; Nitsos and Rees, 1990). The hippocampus and cerebellum were chosen for analysis because both regions show pronounced BDNF and TrkB immunoreactivity during the prenatal period (Dieni and Rees, 2002), and they are also vulnerable to prenatal insults (Dieni and Rees, 2003; Mallard et al., 1999, 2000; Rees and Harding, 1988; Rees et al., 1988).

Materials and methods Uterine artery ligation Chronic placental insufficiency was induced in time-mated, Dunkin–Hartley guinea pigs by unilateral uterine artery ligation as previously described (Lafeber et al., 1984; Nitsos and Rees, 1990). Briefly, pregnant guinea pigs (n = 10) at 28–30 days of gestation (dg, term approximately 67 days) were anesthetized with an intramuscular injection of ketamine (40 mg/kg, Ilium Laboratories, Victoria, Australia) and xylazil (6 mg/kg, Troy Laboratories, Victoria, Australia). A midline incision was made below the umbilicus, the mesometrium associated with one horn of the bicornuate uterus was retrieved, and uterine artery ligation (UAL) was performed at the cervical end of the arterial arcade. Uterine horns were

ligated randomly in each animal to prevent position bias. The ligature remained in place for the duration of the experiment. Fetuses from the unoperated horn served as controls. Tissue preparation At 60 dg, pregnant sows were deeply anesthetized with sodium pentobarbitone (130 mg/kg ip, Nembutal, Merial, Australia), fetuses were removed from the uterine horns, and body weights and crown rump lengths (CRL) recorded. Each fetus was perfusion-fixed in situ via the heart with 0.9% saline solution, followed by 4% paraformaldehyde solution (4% PFA, pH 7.4). Fetal brains were removed and postfixed in 4% PFA for 2 h at 4jC. Blocks of forebrain containing hippocampus and blocks of cerebellar vermis were cryoprotected overnight in 20% sucrose in 0.1 M phosphate buffer (PB, pH 7.4) at 4jC. Coronal serial sections of forebrain (40 m) and parasagittal sections of cerebellum (40 m) were cut on a freezing microtome (Leica, Germany) and collected in 0.1 M PB (pH 7.4). Every fifth section was reserved for Nissl staining for structural analyses. Immunohistochemistry The affinity-purified rabbit anti-BDNF antibodies used in this study were generously donated by Dr. Q. Yan (Amgen, Thousand Oaks, CA) and have previously been characterized in rat (Conner et al., 1996; Yan et al., 1997) and human (Connor et al., 1997). Affinity-purified TrkB antibodies recognizing the intracellular tyrosine domain terminal region (amino acids 794 –804) of full-length TrkB receptor were purchased from Santa Cruz Biotechnology (CA, USA). Both antibodies demonstrate specificity to the respective BDNF and TrkB antigens in the guinea pig brain (Dieni and Rees, 2002). Sections of hippocampus or cerebellum were processed simultaneously for either BDNF or TrkB immunoreactivity (IR). Briefly, sections were washed in tris phosphate-buffered saline (tPBS, pH 7.4) and blocked in 10% bovine serum albumin (BSA, Sigma, USA) in tPBS, followed by incubation with the primary antibodies at 4jC (anti-BDNF, 0.9 Ag/ml, 72 h; anti TrkB, 1:500, 16 h). For negative control staining, the primary antibody was omitted and replaced with normal rabbit IgG (2 Ag/ml, Oncogene Science, Australia). Sections were incubated with secondary antibodies for 1 h at room temperature (biotinylated goat anti-rabbit antibody, 1:200, Vector Laboratories, NSW) and reacted with an avidin –biotin – peroxidase solution (1:100, ABC Elite, Vector Laboratories). The BDNF reaction was visualised with 0.04% nickel-enhanced diaminobenzidene tetrahydrochloride (DAB) solution (Sigma) with 0.3% H2O2 in tPBS, giving a blue –black reaction product. The TrkB reaction was visualized with DAB solution only, giving a brown reaction product. All BDNF- and TrkB-stained sections were counterstained with 0.01% thionine in acetate buffer (pH 4.4). Staining failed to occur when the primary antibodies were replaced with rabbit IgG (not shown).

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

Immunohistochemical analysis Immunostained sections were subjected to semiquantitative analysis, according to the distribution pattern of each antibody. Immunoreactivity (IR) in each hippocampal layer (CA1-4, stratum oriens, stratum radiatum, stratum lacunosum-moleculare, stratum moleculare, stratum granulosum, mossy fiber layer, stratum lucidum, hilar region), and each cerebellar layer (Molecular layer, Purkinje cell layer, internal granule layer, white matter) was assessed according to staining intensity, and in the case of the cellular layers, the pattern of distribution: no staining (0), weak IR (1), moderate IR (2), strong IR (3), intense IR (4). Structural analysis For each animal, five Nissl-stained sections of dorsal hippocampus 200 Am apart were used for measurement of stratum moleculare and stratum granulosum in the dentate gyrus. The mossy fiber layer, consisting of the collateral axons of the dentate granule cells, was measured in adjacent BDNF-stained sections; the layer is clearly defined in these sections. For areal measurement of stratum granulosum and the mossy fiber layer, level-matched sections from each animal were viewed on a projecting microscope (Olympus) at a final magnification of 155. The mean cross-sectional area of each layer was determined using point counting (Gunderson and Jensen, 1987). The mean width of stratum moleculare was determined using Image Pro Plus Software (Media Cybernetics, Maryland). Level-matched sections were viewed under a 10 objective lens, and four evenly spaced measurements were made along the length of the molecular layer overlying the suprapyramidal blade of the dentate gyrus. Cerebellum Nissl-stained, parasagittal sections (400 Am apart) collected through the entire cerebellum were used to determine the volume of the molecular, internal granule, and white matter layers. Sections were viewed on a projecting microscope and the total volume (mm3) of each layer was then determined for each animal using the Cavalieri principle (Gunderson and Jensen, 1987).

able. Significant values were subjected to post hoc analysis using Tukey’s test. A correlation analysis was performed using Spearman’s nonparametric test to determine if there was a significant relationship between BDNF and the growth of the mossy fiber layer and stratum moleculare. Significance was accepted if P < 0.05. All measurements were performed on coded slides.

Results Intrauterine growth restriction Fetuses from the ligated uterine horn were classified as GR using previously established criteria (Nitsos and Rees, 1990). Parameters at 60 dg are summarized in Table 1. The mean body weight of GR fetuses at 60 days (n = 8) was significantly reduced ( P < 0.0001) when compared to controls (n = 7). Individual body weights of GR fetuses at post mortem ranged from 25 to 60 g; fetuses were described as ‘‘severely’’ growth-restricted (GR+) if their body weight was less than 40 g at post mortem. Growthrestricted fetuses also displayed significant reductions in mean crown rump length ( P < 0.0005) and brain ( P < 0.01), cerebellar ( P < 0.01), placental ( P < 0.0005), and liver ( P < 0.0001) weights when compared to controls. The mean brain – liver and brain –body ratios were significantly higher in GR fetuses, when compared to controls, reflecting the relative sparing of the brain compared to the liver ( P < 0.0005) and body ( P < 0.0001), as has been reported previously (Mallard et al., 2000; Nitsos and Rees, 1990; Tolcos and Rees, 1997). There was no evidence of gray or white matter necrosis or infarction in the brains of GR fetuses, in agreement with previous studies of this model (Nitsos and Rees, 1990; Rees et al., 1988, 1997). The effects of CPI on the hippocampus BDNF-IR The staining pattern observed in control hippocampi at 60 dg was comparable to that described in a previous ontogeny Table 1 Body and organ weights

Statistical analysis The mean value of each parameter for control and GR groups was determined by averaging the mean values from each animal. Mean group values of body and organ weights were subjected to a two-sided t test. Mean values of BDNF and TrkB staining intensity were compared using a two-way ANOVA. The effect of ‘‘Treatment’’ (CON or GR) and/or structural layer on staining intensity was determined using ‘‘Treatment’’ and ‘‘Layer’’ as independent variables and mean values of ‘‘staining intensity’’ as the dependent vari-

267

Body (g) CRL (cm) Placenta (g) Liver (g) Brain (g) Cerebellum (g) B/L% B/B%

Control (n = 7)

GR (n = 8)

P

94.49 12.77 5.12 4.53 2.63 0.29 59.2 3.17

43.81 F 9.81 F 3.89 F 1.71 F 2.11 F 0.21 F 138.7 F 5.22 F

**** *** *** **** * * *** ****

F F F F F F F F

1.96 0.45 0.17 0.27 0.04 0.01 3.42 1.70

4.04 0.35 0.155 0.21 0.17 0.04 10.54 0.21

Data expressed as mean F SEM, *P < 0.01, **P < 0.001, ***P < 0.0005, ****P < 0.0001. B/L, B/B: brain – liver, brain – body ratios (expressed as percentages).

268

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

study (Dieni and Rees, 2002). Specifically, a distinct, laminar staining pattern was seen throughout the neuronal layers and their associated process layers, with moderate BDNF-immunoreactivity (-IR) in strata oriens, radiatum, and moleculare (inner third), and intense BDNF-IR in the mossy fiber layer (Fig. 1C). CA1 neurons displayed weak BDNF-IR while

CA2, CA3, and hilar neurons all displayed moderate BDNFIR (not shown). The majority of dentate granule cells were BDNF-positive except for the immature basal cells (Fig. 1D). Semiquantitative analysis of BDNF staining intensity demonstrated that CPI had an overall significant effect on BDNF immunoreactivity ( F = 43.42, P < 0.001, Fig. 1A).

Fig. 1. BDNF-IR (C – F) and TrkB-IR (G – L) in hippocampi from control and growth-restricted fetuses at 60 dg. (A) Quantitation of BDNF-IR in hippocampal layers of control and GR animals. Note significant reductions in BDNF-IR in the neuropil layers. (B) Quantitation of TrkB-IR in hippocampal layers of control and GR animals. Note the significant increases in TrkB-IR in most layers. (C) BDNF-IR in the hippocampus from a representative control fetus. (D) BDNF-IR in dentate gyrus of control fetus. White arrowhead labels basal cells that are BDNF-negative. (E) Reduced BDNF-IR in all layers in the hippocampus from a representative GR+ fetus. (F) BDNF-IR in the dentate gyrus from a GR+ fetus. Note staining is confined to granule cells in the outer half of SG (arrows). (G) TrkB-IR in the hippocampus of a control fetus. (H) TrkB-IR in the control dentate gyrus. (I) Increased TrkB-IR in the hippocampus of GR+ fetus. Note strong staining in the CA1 apical dendrites within SR and SL-M. (J) TrkB-IR in the dentate gyrus of a GR+ fetus. Note that staining is confined to cell bodies in the outer half of SG (arrows). H: hilus; MF: mossy fiber layer; SG: stratum granulosum; SL-M: stratum lacunosum-moleculare; SLu: stratum lucidum; SM: stratum moleculare; SMi: stratum moleculare, inner third; SMo: stratum moleculare, outer two thirds; SO: stratum oriens; SR: stratum radiatum. Scale bars: C, E, G, I = 200 Am; D, F, H, J = 50 Am, *P < 0.05, **P < 0.005, error bars represent SEM.

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

Post hoc analysis revealed overall significant reductions in BDNF-IR in the neuropil layers (strata oriens, radiatum, lucidum, inner moleculare, and mossy fiber layer) within the hippocampi of the total cohort of GR animals. When assessed as a separate cohort, the hippocampi of GR+ fetuses showed a marked reduction in BDNF-IR across all fiber and cellular layers (Fig. 1E). The staining pattern of BDNF-IR in stratum granulosum (Fig. 1F) was similar to the pattern seen at 50 dg in the same region (Dieni and Rees, 2002). TrkB-IR In control hippocampi at 60 dg, strong TrkB-IR is evident in pyramidal neurons, with staining in the apical dendrites (Fig. 1G). The majority of dentate granule neurons is TrkB-positive except for the immature basal cells. There is also weak TrkB-IR along the inner portion of the molecular layer (Fig. 1H). Semiquantitative analysis of TrkB staining intensity revealed that chronic placental insufficiency had an overall significant effect on TrkB immunoreactivity ( F = 56.01, P < 0.001, Fig. 1B). Post hoc analysis demonstrated significant increases in TrkB-IR in all cellular and neuropil layers except the outer two-thirds of stratum moleculare, the mossy fiber layer, and stratum oriens. This is illustrated by comparing Figs. 1I with G . The most marked changes in TrkB-IR were apparent in GR+ fetuses; these hippocampi also displayed a dramatic reduction in BDNF-IR, as described above. In one of the GR+ animals, there was also distinctive TrkB-IR in the apical dendrites of dentate granule cells, which was not observed in hippocampi of any other GR or control animals (not shown). In two other GR+ animals, the outer half of stratum granulosum was TrkB-positive while the inner half was TrkB-negative (Fig. 1J). This staining pattern was similar to that seen in 50 dg hippocampi from controls (Dieni and Rees, 2002).

269

between decreasing intensity of BDNF-IR and a decreasing area of the mossy fiber layer ( P < 0.001) and stratum moleculare ( P < 0.05). The effects of CPI on the cerebellum BDNF-IR Semiquantitative analysis demonstrated that there was no significant effect of CPI on BDNF-IR in any of the cerebellar layers (Fig. 2A). Strong BDNF-IR was observed in Purkinje cell bodies, scattered cell bodies in the internal granule layer and in the neuropil of both the molecular layer and the internal granule layer in control animals (Figs. 2C, E). A similar distribution of BDNF-IR in GR fetuses is illustrated in Figs. 2D and F. TrkB-IR Semiquantitative analysis of TrkB-IR in each cerebellar layer revealed that there was no significant effect of CPI on the distribution or intensity of the protein (Fig. 2B). TrkB-IR was observed in Purkinje cell bodies and dendrites, as well as interneurons in the molecular layer (Fig. 2G). Cells in the premigratory or migratory zones of the external granule layer showed weak TrkB-IR while mature granule cells in the internal granule layer showed strong TrkB-IR. In the internal granule layer, TrkB-IR was more intense in the deep (Fig. 2I) rather than superficial regions (Fig. 2K) of the cerebellar folia, in accord with the timing of granule cell maturation (Altman, 1969). Neurons in the deep cerebellar nuclei were also TrkB-positive (not shown). The comparison of TrkBIR in corresponding cerebellar folia from control (Figs. 2G, I, K) and GR (Figs. 2H, J, L) groups illustrates the lack of difference in the staining pattern. Structural analyses

In the dentate gyrus of GR animals compared to controls, there was a reduction in the mean width of stratum moleculare ( P < 0.05), the area of the mossy fiber layer ( P < 0.05), and the ratio of the mossy fiber layer to stratum granulosum ( P < 0.005), although not in the area of stratum granulosum alone (Table 2). There was a significant correlation (r = 0.77)

When compared to controls, cerebella from GR fetuses were significantly reduced in the volumes of the molecular layer ( P < 0.05), internal granule layer ( P < 0.01), and the white matter ( P < 0.01) (Table 3). This is illustrated by comparing Figs. 2M and P (control) with Figs. 2N and Q (GR). In the cerebella of GR+ fetuses (Fig. 2O), there was a qualitative reduction in the degree of foliation in some lobules, particularly lobules II and IX, when compared to controls.

Table 2 Measurement of hippocampal layers

Discussion

Structural analyses

Control (n = 7) SM (Am) MF (Am2) MF/SG% SG (Am2)

205.83 703.2 95.9 682.2

F F F F

4.38 76.9 2.1 89.6

GR (n = 8) 171.0 470.9 73.8 657.2

F F F F

8.87 51.0 3.8 58.9

P * * ** NS

Data expressed as mean F SEM, *P < 0.05, **P < 0.005, NS: not significant.

This study has unequivocally demonstrated that chronic placental insufficiency, induced via a reduction in uteroplacental blood flow throughout the second half of gestation in the guinea pig affects expression of BDNF and TrkB in the fetal hippocampus and concomitantly causes a reduction in the growth of neural processes in this structure. Although

270

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

Fig. 2. BDNF-IR (A, C – F) and TrkB-IR (B, G – L) in cerebella from control and growth-restricted fetuses at 60 dg. (A) Quantitation of BDNF-IR in cerebellar layers of control and GR animals. (B) Quantitation of TrkB-IR in cerebellar layers of control and GR animals. (C) BDNF-IR in the cerebellum from a control fetus, showing lobules IV – VIII. (D) BDNF-IR in the cerebellum of a GR fetus. Note that there is no change in BDNF-IR. This is further demonstrated by comparing lobule VIII from control (E) and growth-restricted fetuses (F). (G) TrkB-IR in the cerebellum of a control fetus. (H) TrkB-IR in the cerebellum of a growth-restricted fetus. (I, J) Deep regions of the folia show strong TrkB-IR in the ML, PCL, and IGL in control (I) and GR (J). fetuses. Arrows indicate Purkinje cell bodies, arrowheads label interneurons in the ML. Weaker TrkB-IR is seen in the tips of the folia from control (K) and growth-restricted (L) fetuses. (M – O) Nissl-stained cerebella from control (M), GR (N), and GR+ (O) fetuses. Note alterations in foliation in lobules II and IX in N and lobules II, IV, VI, and IX in O, as indicated by the arrows, compared to controls P. Neuropil and cellular layers in the cerebellum of a control fetus. (Q) Reduced volume of neuropil and cellular layers in the cerebellum of a growth-restricted fetus. DCN: deep cerebellar nuclei; EGL: external granular layer; ML: molecular layer; PCL: Purkinje cell layer; IGL: internal granular layer; WM: white matter. Scale bars: C, D = 1 mm; E, F = 250 Am; G, H = 500 Am, I – L = 50 Am, M – O = 2 mm; P, Q = 100 Am. Error bars represent SEM.

this study was not designed to demonstrate a direct link between the severity of chronic placental insufficiency, neurotrophin levels, and neurite growth, the finding that the most dramatic reduction in BDNF-IR was associated

with marked attenuation of dentate granule cell process growth in the most severely growth-restricted fetuses would support the notion that there is at least an association between these parameters. These findings are significant

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273 Table 3 Volumes of cerebellar layers 3

ML (mm ) IGL (mm3) WM (mm3)

Control (n = 7)

GR (n = 8)

P

46.9 F 1.58 53.04 F 1.74 28.4 F 1.26

27.83 F 3.31 35.69 F 2.0 18.6 F 1.94

* ** **

Data expressed as mean F SEM, *P < 0.05, **P < 0.01.

in that they demonstrate for the first time that chronic, adverse intrauterine conditions can downregulate neurotrophin expression in the fetus and that such effects could contribute to altered brain development in compromised offspring. We note that CPI does not have a global effect on BDNF and TrkB expression in the brain as the levels of these proteins did not appear to be altered in the cerebellum although cerebellar growth was significantly reduced. Possible reasons for this differential effect are explored below. The effects of chronic placental insufficiency on the developing hippocampus We have shown that CPI results in significant reductions in the growth of the molecular and mossy fiber layers in the dentate gyrus of GR fetuses, indicating a reduction in dendrite and axonal growth, respectively. Previously in the model, we have shown that there are also reductions in the growth of stratum oriens and in the number of CA1 pyramidal cells (Mallard et al., 2000). In the present study, we have shown that adverse effects on hippocampal growth were accompanied by a reduction in BDNF expression. However, as indicated above, the design of the study does not allow us to propose more than an association between these factors. The upregulation in TrkB might occur in response to the downregulation of BDNF; this is a typical response of the receptor when the ligand is reduced (Baek et al., 1996; Schaaf et al., 1998). Our model of CPI is known to result in hypoxemia (Jensen et al., 1996), hypoglycemia (Jones and Parer, 1983), and alterations to insulin-like growth factor-1 and thyroid hormone (Jones et al., 1990), as well as BDNF. We consider that intrauterine compromise in guinea pigs and also in humans is likely to include all of these factors to varying degrees, depending on individual circumstances including the physiological status of the fetus. However, it is important to focus on specific factors, as we have done here, in an attempt to identify proteins that might have effects on specific aspects of brain growth, such as axonal and dendritic proliferation. At this stage, it is not possible to determine whether CPI has a direct effect on the production, turnover, or release of BDNF. However, it is known that BDNF synthesis is largely dependent on calcium influx (Du et al., 2000; Tao et al., 1998) and that the maintenance of calcium homeostasis is compromised during hypoxia (Mishra and Delivoria-Papadopolous, 1999; Vannucci, 1990) and hypoglycemia (Mattson et al., 1993). Reduced synthesis of BDNF in dentate granule cells might lead to a decrease in the anterograde transport of

271

the protein along the mossy fiber pathway and affect the survival of post-synaptic CA1 and CA3 neurons; as mentioned above, we have shown that there is a reduction in the number of these neurons in CPI (Mallard et al., 2000). It is relevant to note that a reduction in BDNF-IR in CA1 pyramidal neurons following ischemia has been shown to precede cell death (Yamasaki et al., 1998). Another aspect to consider is that CPI might delay neural development and affect the establishment of the normal distribution pattern of BDNF-IR and TrkB-IR. There is some evidence for this in the finding that the distribution of these two proteins at 60 dg resembles the pattern seen at 50 dg (Dieni and Rees, 2002). Whether this occurs as a direct effect of CPI or secondarily to a delay in cellular development is uncertain. The effect of chronic placental insufficiency on the developing fetal cerebellum In agreement with previous studies of prenatal compromise in the sheep (Rees and Harding, 1988; Rees et al., 1988, 1997) and guinea pig (Mallard et al., 2000), we have shown that the developing cerebellum is susceptible to adverse intrauterine conditions. The volume of the mole cular, internal granule, and white matter layers were reduced compared to controls, reflecting reductions in Purkinje cell dendritic growth, granule cell numbers, and axonal growth or myelination, respectively. Furthermore, the reduced lo bular foliation observed in severely growth-restricted fetuses suggests that the normal pattern of foliation is either delayed or altered by severe chronic placental insufficiency. The latter seems most likely as we have shown in a companion study that altered growth of the cerebellum persists into young adulthood (Emery et al., 2002). In contrast to the hippocampus, the altered cerebellar growth pattern was not associated with alterations in the presence BDNF and TrkB protein expression, at least not at the time point examined in this study. Although BDNF appears to be crucial for some aspects of cerebellar development (Schwartz et al., 1997), CPI might adversely influence other neurotrophins, such as NT-3, which also influence cerebellar growth (Bates et al., 1999; Shalizi et al., 2003). This will be tested in future studies. It should also be noted however that cerebellar neurons synthesize only part of the total cerebellar BDNF, the remainder originating from sources such as the inferior olivary nucleus in the brainstem (Murer et al., 1999; Rocamora et al., 1993), a region that is relatively spared from chronic prenatal insults (Tolcos et al., 2003); this could contribute to the lack of effects on BDNF-IR in the cerebellum after CPI. This study has shown that a chronic reduction in uteroplacental blood flow leads to reduced growth of the fetal hippocampus and that this is associated with alterations in BDNF-IR and TrkB-IR expression. Although we acknow ledge that many factors might be involved in CPI, the finding that the most profound effects on growth occurred

272

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273

in the fetuses where BDNF-IR levels were most markedly reduced suggests that neurotrophic levels are likely to be involved in altered brain growth in compromised intrauterine development. The observation that the reduced growth of the fetal cerebellum was not associated with altered levels of BDNF and TrkB proteins suggests that other agents are likely to be involved. Identifying factors that are affected by intrauterine compromise will ultimately provide a basis from which to attempt therapeutic intervention to redress normal brain development.

Acknowledgments The authors thank Mr. Todd Briscoe for his assistance with the statistical analysis. This project was funded by the National Health and Medical Research Council, Australia.

References Altman, J., 1969. Autoradiographic and histological studies of postnatal neurogenesis. III. Dating the time of production and onset of differentiation of cerebellar microneurons in rats. J. Comp. Neurol. 136, 269 – 294. Baek, J.K., Heaton, M.B., Walker, D.W., 1996. Up-regulation of highaffinity neurotrophin receptor, trk B-like protein on western blots of rat cortex after chronic ethanol treatment. Brain Res. Mol. Brain Res. 40, 161 – 164. Bartrup, J.T., Moorman, J.M., Newberry, N.R., 1997. BDNF enhances neuronal growth and synaptic activity in hippocampal cell cultures. NeuroReport 8, 3791 – 3794. Bates, B., Rios, M., Trumpp, A., Chen, C., Fan, G., Bishop, J.M., Jaenisch, R., 1999. Neurotrophin-3 is required for proper cerebellar development. Nat. Neurosci. 2, 115 – 117. Borghesani, P.R., Peyrin, J.M., Klein, R., Rubin, J., Carter, A.R., Schwartz, P.M., Luster, A., Corfas, G., Segal, R.A., 2002. BDNF stimulates migration of cerebellar granule cells. Development 129, 1435 – 1442. Conner, J.M., Yan, Q., Varon, S., 1996. Distribution of brain-derived neurotrophic factor in the rat pituitary gland. NeuroReport 7, 1737 – 1940. Connor, B., Young, D., Yan, Q., Faull, R.L.M., Synek, B., Dragunow, M., 1997. Brain-derived neurotrophic factor is reduced in Alzheimer’s disease. Mol. Brain Res. 49, 71 – 81. Dieni, S., Rees, S., 2002. Distribution of brain-derived neurotrophic factor and TrkB receptor proteins in the fetal and postnatal hippocampus and cerebellum of the guinea pig. J. Comp. Neurol. 454, 229 – 240. Dieni, S., Rees, S., 2003. Dendritic morphology is altered in hippocampal neurons following prenatal compromise. J. Neurobiol. 55, 41 – 52. Du, J., Feng, L., Yang, F., Lu, B., 2000. Activity- and Ca2+-dependent modulation of surface expression of brain-derived neurotrophic factor receptors in hippocampal neurons. J. Cell Biol. 150, 1422 – 1423. Emery, B., Rehn, A., Rees, S., 2002. Chronic placental insufficiency results in long term alterations to cerebellar circuitry. In: Pilowsky, P. (Ed.), Proceedings of the Australian Neuroscience Society, vol. 13. Australian Neuroscience Society Inc., Sydney, p. 72. Ferrer, I., Lopez, E., Pozas, E., Ballabriga, J., Marti, E., 1998. Multiple neurotrophic signals converge in surviving CA1 neurons of the gerbil hippocampus following transient forebrain ischaemia. J. Comp. Neurol. 394, 416 – 430. Gorski, J.A., Zeiler, S.R., Tamowski, S., Jones, K.R., 2003. Brain-derived neurotrophic factor is required for the maintenance of cortical dendrites. J. Neurosci. 23, 6856 – 6865.

Gunderson, H.J.G., Jensen, E.B., 1987. The efficiency of systematic sampling in stereology and its prediction. J. Microsc. 147, 229 – 263. Ip, N.Y., Li, Y., Yancopoulos, G.D., Lindsay, R.M., 1993. Cultured hippocampal neurons show responses to BDNF, NT-3 and NT-4, but not NGF. J. Neurosci. 18, 3394 – 3405. Jensen, A., Klonne, H.J., Detmer, A., Carter, A.M., 1996. Catecholamine and serotonin concentrations in fetal guinea pig brain: relation to regional cerebral blood flow and oxygen delivery in the growth-restricted fetus. Reprod. Fertil. Dev. 8, 355 – 364. Jones, C.T., Parer, J.T., 1983. The effect of alterations in placental blood flow on the growth of and nutrient supply to the fetal guinea pig. J. Physiol. 343, 525 – 537. Jones, C.T., Lafeber, H.N., Rolph, T.P., Parer, J.T., 1990. Studies on the growth of the fetal guinea pig. The effects of nutritional manipulation on prenatal growth and plasma somatomedin activity and insulin-like growth factor concentrations. J. Dev. Physiol. 13, 189 – 197. Kokaia, Z., Andsberg, G., Yan, Q., Lindvall, O., 1998. Rapid alterations BDNF protein levels in the rat brain after focal ischemia: evidence for increased synthesis and anterograde axonal transport. Exp. Neurol. 154, 289 – 301. Lafeber, H.N., Rolph, T.P., Jones, C.T., 1984. Studies on the growth of the fetal guinea pig. The effects of ligation of the uterine artery on organ growth and development. J. Dev. Physiol. 6, 441 – 459. Larroque, B., Betrais, S., Czernichow, P., Lener, J., 2001. School difficulties in 20-year-olds who were born small for gestational age at term in a regional cohort study. Pediatrics 108, 111 – 115. Low, J.A., Handley-Derry, M.H., O’Burke, S., Peters, R.D., Pater, E.A., Killen, H.L., Derrick, E.J., 1993. Association of intrauterine fetal growth retardation and learning deficits at age 9 – 11 years. Am. J. Obstet. Gynecol. 167, 1499 – 1505. Lowenstein, D.H., Arsenault, L., 1996. The effects of growth factors on the survival and differentiation of cultured dentate granule neurons. J. Neurosci. 16, 1759 – 1769. Mallard, E.C., Rees, S., Stringer, M., Cock, M.L., Harding, R., 1998. Effects of chronic placental insufficiency on brain development in fetal sheep. Pediatr. Res. 43, 262 – 270. Mallard, E.C., Rehn, A., Rees, S., Tolcos, M., Copolov, D., 1999. Ventriculomegaly and reduced hippocampal volume following intrauterine growth restriction: implications for schizophrenia. Schizophr. Res. 40, 11 – 21. Mallard, C., Loeliger, M., Copolov, D., Rees, S., 2000. Reduced numbers of neurons in the hippocampus and cerebellum in the postnatal guinea-pig following intrauterine growth restriction. Neuroscience 100, 327 – 333. Mattson, M.P., Zhang, Y., Bose, S., 1993. Growth factors prevent mitochondrial dysfunction, loss of calcium homeostasis, and cell injury, but not ATP depletion in hippocampal neurons deprived of glucose. Exp. Neurol. 121, 1 – 13. McAllister, A.K., 2001. Neurotrophins and neuronal differentiation in the central nervous system. Cell Mol. Life Sci. 58, 1054 – 1060. McIntire, D.D., Bloom, S.L., Casey, B.M., Leveno, K.J., 1999. Birth weight in relation to morbidity and mortality among newborn infants. N. Engl. J. Med. 340, 1234 – 1238. Merlio, J.-P., Ernfors, P., Kokaia, Z., Middlemas, D.S., Bengzon, J., Kokaia, M., Smith, M.-L., Siesjo, B.K., Hunter, T., Lindvall, O., Persson, H., 1993. Increased production of TrkB protein tyrosine kinase receptor after brain insults. Neuron 10, 151 – 164. Mishra, O.P., Delivoria-Papadopolous, M., 1999. Cellular mechanisms of hypoxic injury in the developing brain. Brain Res. Bull. 48, 223 – 238. Murer, M.G., Boissiere, F., Yan, Q., Hunot, S., Villares, J., Faucheux, B., Agid, Y., Hirsch, E., Raisman-Vozari, R., 1999. An immunohistochemical study of the distribution of brain-derived neurotrophic factor in the adult human brain, with particular reference to Alzheimer’s disease. Neuroscience 88, 1015 – 1032. Nitsos, I., Rees, S., 1990. The effects of intrauterine growth retardation on the development of neuroglia in the fetal guinea pigs. An immu-

S. Dieni, S. Rees / Experimental Neurology 192 (2005) 265–273 nohistochemical and ultrastructural study. Int. J. Dev. Neurosci. 8, 233 – 244. Rees, S., Harding, R., 1988. The effects of intrauterine growth retardation on the development of the Purkinje cell dendritic tree in the cerebellar cortex of fetal sheep: a note on the ontogeny of the Purkinje cell. Int. J. Dev. Neurosci. 6, 461 – 469. Rees, S., Bocking, A.D., Harding, R., 1988. Structure of the fetal sheep brain in experimental growth retardation. J. Dev. Physiol. 10, 211 – 224. Rees, S., Stringer, M., Just, Y., Hooper, S.B., Harding, R., 1997. The vulnerability of the fetal brain to hypoxemia at midgestation. Dev. Brain Res. 103, 103 – 118. Rocamora, N., Garcia-Ladona, F.J., Palacios, J.M., Mengod, G., 1993. Differential expression of brain-derived neurotrophic factor, neurotrophin-3, and low-affinity nerve growth factor receptor during the postnatal development of the rat cerebellar system. Mol. Brain Res. 17, 1 – 8. Rudge, J.S., Mather, P.E., Pasnikowski, E.M., Cai, N., Corcoran, T., Acheson, A., Anderson, K., Lindsay, R.M., Weigand, S.J., 1998. Endogenous BDNF protein is increased in adult rat hippocampus after a kainic acid induced excitotoxic insult but exogenous BDNF is not neuroprotective. Exp. Neurol. 149, 398 – 410. Schaaf, M.J., de Jong, J., de Kloet, E.R., Vreugdenhil, E., 1998. Downregulation of BDNF mRNA and protein in the rat hippocampus by corticosterone. Brain Res. 813, 112 – 120. Schwartz, P.M., Borghesani, P.R., Levy, R.L., Pomeroy, S.L., Segal, R.A., 1997. Abnormal cerebellar development and foliation in bdnf / mice reveals a role for neurotrophins in CNS patterning. Neuron 19, 269 – 281. Shalizi, A., Lehtinen, M., Gaudilliere, B., Donovan, N., Han, J., Konishi, Y., Bonni, A., 2003. Characterization of a neurotrophin signaling mechanism that mediates neuron survival in a temporally specific pattern. J. Neurosci. 23, 7326 – 7336.

273

Takeda, A., Onodera, H., Sugimoto, A., Kogure, K., Obinato, M., Shabihara, S., 1993. Coordinated expression of messenger RNAs for nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3 in the rat hippocampus following transient forebrain ischemia. Neuroscience 55, 23 – 31. Tao, X., Finkbeiner, S., Arnold, D.B., Shaywitz, A., Greenberg, M., 1998. Ca2+ influx regulates bdnf transcription by a CREB family transcription factor-dependent mechanism. Neuron 20, 709 – 726. Tolcos, M., Rees, S., 1997. Chronic placental insufficiency in the fetal guinea pig affects neurochemical and neuroglial development but not neuronal numbers in the brainstem: a new method for combined stereology and immunohistochemistry. J. Comp. Neurol. 379, 99 – 112. Tolcos, M., Harding, R., Loeliger, M., Breen, S., Cock, M., Duncan, J., Rees, S., 2003. The fetal brainstem is relatively spared from injury following intrauterine hypoxemia. Brain Res. Dev. Brain Res. 143, 73 – 81. Vannucci, R.C., 1990. Experimental biology of cerebral hypoxia – ischemia: relation to perinatal brain damage. Pediatr. Res. 27, 317 – 326. Walton, M., Connor, B., Lawlor, P., Young, D., Sirimanne, E., Gluckmann, P., Cole, G., Draganow, M., 1999. Neuronal death and survival in two models of hypoxic – ischemic brain damage. Brain Res. Rev. 29, 137 – 168. Yamasaki, Y., Shigeno, T., Furukawa, Y., Furukawa, S., 1998. Reduction in brain-derived neurotrophic factor protein level in the hippocampal dendritic field precedes delayed neuronal damage in the rat brain. J. Neurosci. Res. 53, 318 – 329. Yan, Q., Radeke, M.J., Matheson, C.R., Talvenheimo, J., Welcher, A.A., Feinstein, S.C., 1997. Immunocytochemical localisation of TrkB in the central nervous system of the adult. J. Comp. Neurol. 378, 135 – 157.